Academic journal article Administrative Science Quarterly

Why Do Employers Only Reward Extreme Performance? Examining the Relationships among Performance, Pay, and Turnover

Academic journal article Administrative Science Quarterly

Why Do Employers Only Reward Extreme Performance? Examining the Relationships among Performance, Pay, and Turnover

Article excerpt

This study develops an efficiency explanation for commonly observed performance-based compensation contracts that aggressively reward extreme performance while largely disregarding performance distinctions for moderate performance levels. In response to this reward-the-extremes contract, the paper predicts a relationship between performance and turnover that fluctuates by performance level. Evidence of the hypothesized contract and the resulting pattern of turnover are provided empirically with data collected from 984 engineering employees of two large high-technology companies in the San Francisco Bay Area. The data confirm that extremely high and moderately low performers are likely to remain in firms offering these contracts while moderately high and extremely low performers are likely to depart.(*)

Numerous disciplines provide theoretical and empirical support for the efficacy of reward schemes that pay for performance. Social psychologists using experimental methods have found that perfromance-based pay serves to enhacne effort and upgrade workforce quality (Landau and Leventhal, 1976; Leventhal, 1976). Economic theorists have also argued that performance-contingent rewards motivate effort and attract talent when effort and talent are not easily observed (Holmstrom, 1979; Levinthal, 1988). Organization scholars have cited

similar benefits of distributing rewards based on performance (Lawler, 1981; Milkovich and Wigdor, 1991). Consistent with achieving these objectives, the majority of organizations claim to use some form of performance-based pay, particularly for execmpt employees (Peck, 1984; Milkovich and Wigdor, 1991).

However, pay-for performance schemes are typically not the simply incentive contracts or piece rates assumed and examined by expectancy and agency theorists, in which pay increases directly with incremental gains in objectively measured performance. Instead, these commonly used merit-pay schemes subjectively assess performance and largely restrict variation in assessments to three to five categories. Further, only a small percentage are categorized outside one or two large central performance categories (Medoff and Abraham, 1980; Gellerman and Hodgson, 1988; Bretz and Milkovich, 1989; Milkovich and Wigdor, 1991: 84). Although some scholars (e.g., Abbot and Schuster, 1984) advocate forced distributions or ranking procedures to curb this clustering of ratings, in practice such schemes appear rather infrequently (Peck, 1984). Further, other scholars (Gellerman and Hodgson, 1988; Lawler, 1990) explicitly advocate merit-pay schemes that place 80-90 percent of employees within a single performance category.

Pay differences based on performance reflect this limited dispersion in performance ratings. For the majority of employees, the typical perfomance-based contract provides at most one performance-based pay distinction and commonly no distinction whatsoever. Only the small percentage of employees rated in the extreme categories are systematically granted increases substantially higher or lower than those granted to employees in these central categories. Not surprisingly, empirical examinations of these pay schemes confirm rather weak links between pay and performance (Meyer, Kay, and French, 1965; Lawler, 1971; Medoff and Abraham, 1980). Teel (1986), in a study of the pay practices of 16 firms, found that 80 percent of yearly individual salary increases were clustered within 2 percent of the mean increase. Studies by compensation consultants suggest that the link between performance and pay as perceived by employees is also quite weak (HayGroup, Inc. 1989). Hence, although the majority of firms claim to use pay for performance, the actual link between performance and pay is greatly limited for all but the extremes of the performance distribution.

Why do firms adopt such reward schemes? Why don't employers either aggressively reward all performance levels or simply disregard performance distinctions altogether? …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.